LLM Version

[gioelemo]

This PR includes the following:

• Use custom starting position and starting angle for the Agent
• Slightly adapted visualization script to include more than one Agent (passed as additional information).
• Minor changes for LLM version (see PR in terra-baselines)

The executables

DATASET_PATH=/path/to/dataset DATASET_SIZE=<num_maps_per_type> python -m terra.viz.main_manual

and

DATASET_PATH=/path/to/dataset DATASET_SIZE=<num_maps_per_type> python -m terra.viz.play

should work as before

[Copilot]

Pull Request Overview

This PR adds support for custom agent positioning and angles to the Terra environment for LLM integration. It modifies the environment to accept optional custom starting positions and angles for agents while maintaining backward compatibility with existing behavior.

• Support for custom agent starting positions and angles throughout the environment pipeline
• Enhanced visualization system to handle multiple agents with additional information
• Configuration updates to support different map paths and reduced curriculum levels

Reviewed Changes

Copilot reviewed 6 out of 7 changed files in this pull request and generated 6 comments.

Show a summary per file

File	Description
terra/viz/main_manual.py	Adds custom position and angle parameters to environment reset
terra/viz/game/game.py	Extends visualization to support multiple agents and partition rendering
terra/state.py	Passes custom position/angle parameters through state initialization
terra/env.py	Updates environment reset to accept and handle custom agent parameters
terra/config.py	Modifies curriculum configuration paths and comments out unused levels
terra/agent.py	Implements custom positioning logic with validation and fallback mechanisms